What Are Adventitious Agents in Biomanufacturing?

Adventitious agents are unwanted biological contaminants, primarily viruses and mycoplasma, that find their way into pharmaceutical products during manufacturing. They are not supposed to be there, they were not deliberately introduced, and their presence in a vaccine, therapeutic protein, or gene therapy can pose a genuine safety risk. Ensuring their absence is one of the foundational goals of biopharmaceutical safety, and it is harder than it sounds, because the biological raw materials and living cell cultures used to produce these medicines are inherently vulnerable to contamination from sources that are not always obvious.

What Counts as an Adventitious Agent

The term covers any extraneous living or replicating entity that ends up in a biological product without being intended. Viruses get most of the attention because they are small enough to slip through many physical barriers, difficult to culture and identify, and potentially dangerous to patients. But the category also includes mycoplasma (tiny bacteria without cell walls that can silently infest cell cultures), bacteria, fungi, and in some regulatory contexts even prions. The common thread is that none of these belong in the product, and all of them could compromise its safety or its ability to work as intended.

The distinction from “endogenous” agents matters. Many cell lines used in manufacturing carry viral sequences baked into their own genomes. Chinese hamster ovary (CHO) cells, the workhorses of the biologics industry, are known to harbor endogenous retrovirus sequences and to release retroviral-like particles. One research group identified 173 type-C endogenous retrovirus sequences in the CHO genome, clustered into three groups, with one group producing full-length transcripts carrying intact coding regions and loading viral genome RNA into particles that could theoretically be functional.¹ These endogenous retroviruses are a known, ongoing concern, but they are distinct from adventitious agents because they come with the cell line itself rather than sneaking in from outside.

How Contamination Gets In

The raw materials used to grow cells are the single biggest entry point. Fetal bovine serum, a common cell-culture supplement, is a well-documented vector. Bovine viral diarrhea virus and bovine parvovirus have both been traced to contaminated serum lots used during manufacturing.² Serum is drawn from animals, pooled, and processed, and any virus circulating in the donor herd can hitch a ride. The same logic applies to other animal-derived reagents: trypsin from porcine pancreas, for instance, has historically been a source of porcine viruses.

The cell substrates themselves are another route. Even when a cell line is supposedly “clean,” it may harbor viruses picked up during its original establishment decades ago, or acquired through contaminated reagents used in its early passages. Insect cell lines used for recombinant protein expression have been found to be persistently infected with adventitious viruses, raising safety concerns about biologicals produced with those cells.³ Rodent cell lines present their own profile: a study of 185 manufacturer cell banks from mouse, rat, and hamster origins found infectious retrovirus in mouse myeloma and hybridoma lines, though not in hamster or rat lines.⁴

Then there is the manufacturing environment itself: the facility’s air handling, the operators, the water supply, and anything that touches the product during production. A contamination event can trace back to something as mundane as a gasket failure on a bioreactor or a technician carrying a common human virus. The diversity of possible entry points is part of what makes adventitious agent control so challenging. Each combination of product type and contaminant requires its own prevention and detection strategy.

The Move Away from Animal-Derived Materials

One straightforward way to reduce the risk is to stop using the materials that carry it. The industry has been steadily shifting toward chemically defined, serum-free, and xeno-free (animal-component-free) media for cell culture. These formulations replace animal-derived supplements with recombinant proteins, synthetic lipids, and defined chemical mixtures. Beyond reducing adventitious agent risk, they also eliminate the batch-to-batch variability that comes with biological materials and lower costs in some cases.⁵

The transition is not complete, though. Some cell types grow poorly or behave differently without serum, and certain legacy manufacturing processes were validated with animal-derived reagents. Switching reagents means revalidating the entire production process, which is expensive and time-consuming. So while the direction of travel is clear, animal-derived materials have not disappeared from biomanufacturing.

Lessons from Historical Contamination Events

The field’s caution about adventitious agents is not theoretical. Real contamination events have shaped modern testing requirements. The best-known case involves simian virus 40 (SV40), which contaminated poliovirus vaccines administered between 1955 and 1963. The virus was present in the monkey kidney cells used to grow the vaccine and was not detected by the testing methods available at the time. Millions of people received contaminated vaccine, triggering decades of concern about potential cancer links. A large epidemiological analysis found that, compared with unexposed cohorts, those exposed as infants or children showed no increased risk of ependymoma, osteosarcoma, or brain cancers overall, and mesothelioma rates were not significantly elevated either, though the exposed cohorts had not yet reached the peak age for mesothelioma at the time of the study.⁶ The reassuring cancer data did not erase the fundamental lesson: conventional testing missed a contaminant that was there all along.

More recently, porcine circovirus type 1 (PCV1) DNA was detected in a widely used rotavirus vaccine. A retrospective analysis of infants who received the vaccine found an anti-PCV1 antibody seropositivity rate of roughly 1% after two doses, compared with about 0.3% in the placebo group, a difference that was not statistically meaningful.⁷ PCV1 is not known to cause disease in humans, but its detection prompted regulatory action and underscored that contamination can happen even in modern, well-regulated manufacturing.

In the biotherapeutics space, Vesivirus 2117 caused contamination events in multiple biologics manufacturing facilities. Researchers identified and quantified the virus in bioreactor fluids from infected CHO cell cultures using mass spectrometry, finding fragments of two structural proteins and confirming the identification by RT-PCR.⁸ Follow-up work showed that Vesivirus 2117 could infect CHO-K1, BHK-21, and MDCK cell lines, with CHO-K1 cultures showing infection even at very low inoculation levels.⁹ That kind of sensitivity to low-level contamination is exactly what makes adventitious viruses so dangerous in a manufacturing setting: a tiny initial inoculum can amplify quickly in a bioreactor full of susceptible cells.

Traditional Detection Methods and Their Limits

For decades, the standard approach to adventitious virus testing involved two pillars: in vitro assays (exposing indicator cell lines to the test sample and watching for signs of viral damage) and in vivo assays (injecting the sample into animals and monitoring for illness). These methods were established in the mid-twentieth century, and until recently they had never been formally validated against the performance standards that newer methods must meet.

A systematic evaluation tested sixteen viruses against both methods. In vitro cell culture detected all sixteen, though bovine viral diarrhea virus required special conditions and rubella virus was detected with low sensitivity. Many viruses were picked up at levels below one infectious unit. In vivo animal testing performed worse, detecting only six of eleven viruses tested, and four of those required viral doses at least ten times higher than one infectious unit before animals showed signs of disease.¹⁰ Only influenza virus and vesicular stomatitis virus were detected at lower amounts in animals than in cell culture. These findings, combined with ethical pressure to reduce animal use, have pushed the field toward retiring in vivo assays in favor of improved in vitro and molecular methods.

One strength of classical in vitro assays is that they can catch emerging threats that nobody specifically designed a test for. Schmallenberg virus, a cattle and sheep pathogen that emerged in Europe around 2011, was shown to be detectable by standard in vitro assays using Vero, BHK-21, and CHO-K1 detector cells within typical assay timeframes.¹¹ That kind of breadth, the ability to detect something you were not specifically looking for, remains valuable even as newer technologies come online.

Next-Generation Sequencing as a Detection Tool

The most significant shift in adventitious agent detection in recent years has been the adoption of next-generation sequencing. Rather than relying on a virus to kill cells or sicken an animal, NGS reads all the nucleic acid in a sample and computationally identifies sequences that do not belong to the expected host. This approach has already caught viruses that traditional assays missed entirely.¹²

The technology is powerful but comes with practical challenges. Speed is one: traditional Illumina-based sequencing workflows take days. Nanopore-based sequencing offers a faster alternative, and researchers have developed workflows using this platform specifically for comprehensive adventitious virus testing in banked cell lines used in biologics production.¹³ On the bioinformatics side, rapid classification tools using k-mer-based approaches have shown sensitivity comparable to full-read alignment methods while requiring less time and computational memory, though their specificity still needs improvement.¹⁴

Regulatory frameworks are still catching up. An international conference on NGS for adventitious virus detection identified several bottlenecks: the lack of standard reference materials for benchmarking, incomplete and poorly annotated sequence databases, the need for massive data storage and transfer infrastructure, a shortage of bioinformatics expertise, and the absence of harmonized international regulations for using NGS in biologics testing.¹⁵ Until those gaps are closed, NGS will likely serve as a complement to traditional assays rather than a full replacement.

Detecting Mycoplasma

Mycoplasma deserves separate mention because it is not a virus but behaves like a stealth contaminant in similar ways. These tiny bacteria lack cell walls, making them resistant to many common antibiotics and invisible under standard microscopy. They can infest cell cultures without killing cells or producing obvious changes, silently altering cell behavior, gene expression, and experimental results for weeks or months before anyone notices.

Traditional mycoplasma detection relies on culture-based methods that take up to 28 days to return results, an unacceptable delay for fast-moving manufacturing timelines, especially in cell and gene therapy where patient-specific products have short shelf lives. Nucleic acid amplification methods have dramatically shortened that window. One validated assay using standard PCR amplification detects Mycoplasma arginini at fifteen genome copies per reaction and Mycoplasma hominis at one hundred genome copies per reaction, both well below the regulatory thresholds.¹⁶ Even faster systems combining isothermal amplification with CRISPR-based detection can identify mycoplasma specimens down to 0.1 copies per microliter in under an hour.¹⁷

Clearing Viruses from the Final Product

Detection is only half the strategy. No matter how carefully you control raw materials and test cell banks, the possibility of contamination can never be reduced to zero. So biopharmaceutical manufacturing builds in dedicated viral clearance steps: processes designed to physically remove or chemically destroy any virus that might be present.

For monoclonal antibodies and other protein therapeutics, the standard downstream process typically includes at least two independent viral clearance steps, each validated to achieve a certain log reduction value (the number of logs by which the virus titer drops). A reduction of four logs means the step removes 99.99% of the virus present. Regulators expect the combined clearance across all steps to provide a wide safety margin.

Nanofiltration is one of the workhorse methods, using filters with pore sizes in the 15 to 20 nanometer range to physically trap virus particles. Two decades of industry experience show that viruses in the 25 to 40 nanometer range, including picornaviruses, caliciviruses, and SV40, are effectively removed by these filters. When nanofiltration is specifically designed to target small parvoviruses, log reduction factors of four or greater are achievable. When the step is aimed at general viral safety rather than parvovirus retention specifically, parvovirus clearance tends to be lower and more variable.¹⁸

The mechanics of virus filtration are less straightforward than pushing fluid through a sieve. Studies using laser scanning microscopy to visualize virus particles inside filter membranes have revealed that viruses accumulate at specific depths within the filter’s separation layers. When flow is interrupted and then resumed, as can happen during process pauses, viruses can mobilize deeper into the membrane and break through into the filtrate. In one filter type, localized clusters of breakthrough points were responsible for substantial virus transmission.¹⁹ The practical implication is that filter validation must account for real-world process conditions, not just steady-state flow. The same research found that using bacteriophages or nanoparticles as stand-ins for actual parvoviruses during validation has limited applicability, because these surrogates do not behave the same way inside the filter structure.²⁰

Chemical inactivation provides a complementary mechanism. Solvent-detergent treatment, which disrupts the lipid envelopes of viruses, is effective against enveloped viruses like retroviruses and herpesviruses but does nothing to non-enveloped viruses like parvoviruses. Researchers have explored integrating solvent-detergent inactivation directly into chromatography steps, performing it on-column during ion-exchange chromatography to intensify the process. Data suggest effective inactivation of enveloped viruses even at a tenth of the standard solvent-detergent concentration when applied this way.²¹ For tougher targets like polyomaviruses, which are small and non-enveloped, low pH inactivation is an option but generally requires a pH at or below 3 and elevated temperatures to be effective. Polyomaviruses are more resistant to UV-C irradiation than parvoviruses and caliciviruses, making UV treatment unreliable for this class.²²

For continuous chromatography, where product flows through multiple columns in sequence rather than in discrete batches, manufacturers have needed to confirm that viral clearance performance carries over from the traditional single-column format. Testing has shown that virus clearance measured in single-column and multicolumn modes produces similar log reduction values, and that the critical parameters identified in single-column validation are predictive of multicolumn performance.²³

Cell and Gene Therapy Brings New Headaches

The rise of cell and gene therapies has introduced adventitious agent challenges that the traditional biologics industry was never built to handle. These products often use patient-derived cells that are processed ex vivo, genetically modified with viral vectors, and then returned to the patient. The manufacturing window is short, sometimes just days. Many of the harsh viral clearance steps used for protein therapeutics, like low-pH holds or nanofiltration, would damage or destroy the product. You cannot filter a living cell through a 20-nanometer membrane.

The number of approved cell and gene therapy products continues to climb each year, but contamination risks during aseptic processing remain a central concern.²⁴ Quality control for these products includes testing for sterility, adventitious agents, endotoxins, and mycoplasma.²⁵ But the tight timelines mean that results from slower tests may not come back before the product must be administered, especially for autologous therapies made from a patient’s own cells. This places enormous weight on rapid detection methods and on upstream prevention, keeping contaminants out of the process in the first place rather than relying on downstream removal.

Continuous Bioprocessing and the Sampling Problem

The broader push in biomanufacturing toward continuous processes, where cells grow in perfusion bioreactors and product is harvested continuously rather than in discrete batches, creates a distinct challenge for adventitious agent control. In a traditional batch process, if contamination is detected, you discard that batch. In continuous manufacturing, product from an infected bioreactor may have been flowing into downstream processing for hours or days before the contamination is caught.

Modeling work has explored how sampling frequency and sample volume affect the ability to catch contamination in a perfusion bioreactor. For a scenario involving mouse minute virus infecting a high-density CHO cell culture, a noticeable decline in viable cell density did not appear until around day four of infection. To ensure that final products meet a safety standard of 99.999% virus-free, the model indicated that total downstream clearance would need to achieve a log reduction value greater than fifteen, a staggeringly high bar that reflects the enormous viral load that can build up in a high-density continuous culture before contamination is even apparent.²⁶ The implication is that continuous manufacturing may need more frequent in-process monitoring, faster detection tools, and higher validated clearance than batch processes to achieve equivalent safety.

The industry is still working through these questions. Continuous bioprocessing offers real advantages in efficiency and facility footprint, and it is not going away. But the adventitious agent risk profile is fundamentally different from batch manufacturing, and the control strategies are still maturing. This is an area where rapid sequencing-based detection, if it can be made routine and fast enough, could prove transformative: catching a contamination event hours after it begins rather than days later when the bioreactor is already teeming with virus.